Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative

Journal of Bioscience and Bioengineering VOL. xxx No. xxx, xxx, xxxx www.elsevier.com/locate/jbiosc

Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative Makoto Imura,1, 2 Katsuaki Nitta,2 Ryo Iwakiri,1 Fumio Matsuda,3 Hiroshi Shimizu,3 and Eiichiro Fukusaki2, * Mitsubishi Corporation Life Sciences Limited, 1-6 Higashihama, Saiki, Oita 876-8580, Japan,1 Department of Biotechnology, Graduate School of Engineering, Osaka University, 2-1 Yamadaoka, Suita, Osaka 565-0871, Japan,2 and Department of Bioinformatic Engineering, Graduate School of Information Science and Technology, Osaka University, 1-5 Yamadaoka, Suita, Osaka 565-0871, Japan3 Received 1 May 2019; accepted 29 July 2019 Available online xxx

The Crabtree effect involves energy management in which yeasts utilize glycolysis as the terminal electron acceptor instead of oxygen, despite the presence of sufficient dissolved oxygen, when oxygen concentrations exceed a certain limit. The Crabtree effect is detrimental to bakery yeast production, because it results in lower cellular glucose yields. Batch culture of Saccharomyces cerevisiae, a Crabtree positive yeast, decreased the cell yield of glucose and produced large amounts of ethanol despite a high specific glucose consumption rate compared to Candida utilis, a Crabtree negative yeast. This study investigated the effect of these characteristics on metabolite levels. We performed metabolome analysis of both yeasts during each growth phase of batch culture using liquid chromatography-tandem mass spectrometry and gas chromatography-mass spectrometry. Principle component analysis of metabolome data indicated that the Crabtree effect affected metabolites related to NADH synthesis in central metabolism. The amount of these metabolites in S. cerevisiae was lower than that in C. utilis. However, to maintain the specific glucose consumption rate at high levels, yeasts must avoid depletion of NADD, which is essential for glucose utilization. Our results indicated that NADH was oxidized by converting acetaldehyde to ethanol in S. cerevisiae, which is in accordance with previous reports. Therefore, the specific NADH production rates of S. cerevisiae and C. utilis did not show a difference. This study suggested that NADD/NADH ratio is disrupted by the Crabtree effect, which in turn influenced central metabolism and that S. cerevisiae maintained the NADD/NADH ratio by producing ethanol. Ó 2019, The Society for Biotechnology, Japan. All rights reserved. [Key words: Metabolomics; The Crabtree effect; NADþ/NADH; Batch culture; Saccharomyces cerevisiae; Candida utilis]

The Crabtree effect involves energy management in which yeasts utilize glycolysis as the terminal electron acceptor instead of oxygen, despite sufficiently high dissolved oxygen concentration content, when such concentrations exceed a certain limit (1). This effect is especially evident in bakery yeast. When respiration is inhibited, Saccharomyces cerevisiae produces ethanol and a low cell yield of glucose. However, the detailed mechanism underlying this effect remains unclear (2). During the manufacturing process, the glucose supply in fed batch or continuous culture should be constrained to inhibit the Crabtree effect, though these cultures result in low daily output. Although previous studies disrupted the activity of pyruvate decarboxylase, which converts pyruvate to acetaldehyde in S. cerevisiae, the disruptant was unable to grow on synthetic media that used glucose as the sole carbon source (3). Therefore, the Crabtree effect appears to be an important phenomenon in aerobically grown S. cerevisiae, although it is

* Corresponding author. Tel./fax: þ81 6 6879 7424. E-mail addresses: [email protected] (M. Imura), katsuaki_nitta@ bio.eng.osaka-u.ac.jp (K. Nitta), [email protected] (R. Iwakiri), fmatsuda@ ist.osaka-u.ac.jp (F. Matsuda), [email protected] (H. Shimizu), fukusaki@ bio.eng.osaka-u.ac.jp (E. Fukusaki).

disadvantageous in terms of growth and cell yield if used in a manufacturing process. On the other hand, certain yeasts do not exhibit the Crabtree effect and consequently, these yeasts may achieve higher cell yields in aerobic culture than Crabtree positive yeasts (4e6). Candida utilis is known to be a Crabtree negative yeast and the Food and Drug Administration recognizes that C. utilis as well as S. cerevisiae are harmless as food additives (7). Therefore, C. utilis is utilized for commercial production of glutathione and yeast extract involving recently developed large scale cultural techniques. Moreover, recombinant C. utilis has been reported to produce lactate (8,9) and isopropanol in laboratory experiments (10). However, due to high ploidy, genetic recombination of C. utilis has not been thoroughly developed (11e13). Although the high cell yield achieved by Crabtree negative yeasts is commercially important, the underlying metabolic mechanisms remain unclear. Metabolomics pertain to the comprehensive analysis of metabolites based on genomic information. It is utilized for precise phenotype analysis since the amount of metabolites is closely associated with phenotype. Metabolomics provides a better understanding of the metabolic machinery affecting a phenotype by comparing the amount of metabolites derived from different phenotypes (14). In our previous studies, metabolomics helped to improve tolerance of 1-butanol (15) and ethanol (16) and better

1389-1723/$ e see front matter Ó 2019, The Society for Biotechnology, Japan. All rights reserved. https://doi.org/10.1016/j.jbiosc.2019.07.007

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

2

IMURA ET AL.

J. BIOSCI. BIOENG.,

understand the characteristics and metabolite levels of the genus Jeotgalicoccus, which possess different capacities to produce terminal alkenes (17). Therefore, analysis of metabolomics results would expectedly reveal changes occurring in metabolite levels of phenotypes induced by the Crabtree effect. In this study, aerobic batch cultures were performed in S. cerevisiae and C. utilis. Metabolomics of cells in each growth phase were conducted. In addition, we compared specific NADH production and consumption rates according to culture results. Based on these results, we discussed the influence of low cell yield and ethanol production induced by the Crabtree effect on metabolite levels.

MATERIALS AND METHODS Strain and cultivation S. cerevisiae NBRC101557 and C. utilis NBRC0988 were purchased from the Biological Resource Center, NITE (Chiba, Japan). Pre-cultivation was performed at 30 C and 180 rpm (BF-43FL MR, TAITEC Corporation, Saitama, Japan) for 18 h in flask (Hario Co., Ltd., Tokyo, Japan) including 100 mL of YPD medium (10 g/L dried yeast extract, 20 g/L Hipolypepton, and 20 g/L D-glucose). Pre-culture broth was inoculated into batch culture so that initial density was OD600 ¼ 0.1. The synthetic medium used, contained, 10 g/L D-glucose, 1.5 g/L KH2PO4, 0.5 g/L MgSO4$7H2O, 0.06 g/L CaCl2, 5 g/L (NH4)2SO4, 0.4 g/L K2SO4, 0.1 mg/L biotin, 1.5 mg/L D-pantothenic acid hemicalcium salt, 60 mg/L myoinositol, 3 mg/L pyridoxine hydrochloride, 14 mg/L thiamine hydrochloride, 0.2 mg/L CuSO4$5H2O, 4 mg/L ZnSO4$7H2O, 10 mg/L FeSO4$7H2O, and 30 mg/L Biospumex 36K (Cognis, Manheim, Germany). D-glucose-1-13C was used for flux ratio analysis, instead of D-glucose. All above chemicals were from Fujifilm Wako Pure Chemical Corporation (Osaka, Japan), or Sigma (St. Louis, MO, USA) unless otherwise indicated. For metabolomics, batch culture was performed in a 10-L jar fermenter (Mitsuwa Frontech, Osaka, Japan) with a working volume of 5 L at 30 C. Next, 4 M NaOH was added to maintain pH at 5.0. Aeration rate and agitation speed were maintained at 1.0 L/min and 700 rpm, respectively. For flux analysis, batch culture was performed in Bio Jr. 8 (ABLE Corporation, Tokyo, Japan) with a working volume of 100 mL at 30 C. Additionally pH was maintained at 5.0 by adding 1 M NaOH. Aeration rate and agitation speed were maintained at 1.0 L/ min and 1000 rpm, respectively. Analysis of biomass and ethanol concentration OD600 was measured using a Double beam Spectrophotometer U-2900 (Hitachi High-Technologies Corporation, Tokyo, Japan). Cell dry mass (CDW) was estimated using OD600 ¼ 1 to 0.27 gCDW/L (S. cerevisiae) and 0.29 gCDW/L (C. utilis). To measure the concentrations of acetate, citrate and succinate in the culture supernatant, an Organic Acid Analysis System (Shimadzu Corporation, Kyoto, Japan) was used. BF96ASX (Oji Scientific Instrument Co., Ltd., Hyogo, Japan) was used to measure glycerol and ethanol concentrations, while Autokit Glucose (Fujifilm Wako Pure Chemical Corporation) was used to measure glucose concentration. Optical dissolved oxygen (DO) sensor InPro 6860i (Mettler-Toledo International Inc., Columbus, OH, USA) was used to measure DO. DO at the start of batch culture was defined as 100% and each DO during growth represented the relative value. Calculation of specific rates Specific rates were calculated from the concentration of cell, glucose and organic acids by Eqs. 1e3. dX/dt ¼ mX

(1)

dP/dt ¼ rX

(2)

dS/dt ¼ nX

(3)

X, P, and S represent cell, product and substrate concentration, respectively. m, r and n represent specific growth rate, specific production rate and specific consumption rate, respectively. t represents time. m, r and n were determined at growth phases when all specific production and consumption rates were constant. Sample preparation for metabolomics Cell collection and metabolite extraction were performed as described previously (18), with minor modifications. The OD600 during growth was measured and an appropriate volume of culture was rapidly transferred into a mixed cellulose ester membrane filter with a 47 mm diameter and 0.45 mm pore size (Toyo Roshi Kaisha, Ltd., Tokyo, Japan) via vacuum filtration and washed with distilled water. The volume of the collected culture broth was adjusted based on the OD600 using the following formula: sampling volume (mL)  OD600 ¼ 80. After the filter paper was lyophilized overnight, 5.0 mg of cells was measured gravimetrically, and 90 mL of internal standard (0.2 mg/mL Ribitol for gas chromatography-mass spectrometry (GCeMS), 0.2 mg/ mL (1S)-(þ) 10 camphorsulfonic acid for liquid chromatography-tandem mass spectrometry (LC-MS/MS)) was added. Next, 1.0 mL of mixed solvent (methanol/

H2O/chloroform, 5/2/2 [v/v/v]) was added and vortexed (Vortex-Genie 2; Scientific Industries, Inc., Bohemia, NY, USA) for 20 s. Metabolites were extracted from the cells using Thermomixer comfort (Eppendorf Co., Ltd., Hamburg, Germany) at 37 C and 1200 rpm for 30 min. Following centrifugation at 4 C and 16,000 g for 3 min, 900 mL of solution was transferred to a new tube, and 450 mL of Ultrapure Water (Fujifilm Wako Pure Chemical Corporation) was added before vortexing for 20 s. After centrifugation at 4 C and 16,000 g for 3 min, 700 mL and 350 mL of the upper polar phase were transferred to new tubes for GCeMS and LC-MS/MS, respectively. The samples were dried via centrifugation for 2 h and lyophilized overnight. GC/MS analysis Hundred microliters of methoxyamine hydrochloride (20 mg/mL-pyridine, Fujifilm Wako Pure Chemical Corporation) was mixed with the lyophilized sample and incubated at 30 C and 1200 rpm for 90 min using a Thermomixer comfort. Next, 50 mL of N-methyl-N-(trimethylsilyl)trifluoroacetamide (GL Sciences Inc., Kyoto, Japan) was added and the preparation was incubated in the Thermomixer comfort at 30 C and 1200 rpm for 30 min. GC/MS analysis was performed as described previously (19). The analysis employed a GC/MS-QP2010Ultra (Shimadzu) and GCMS solution ver. 4.20b software (Shimadzu) to acquire GC and MS data. A 30 m  0.25 mm i.d. DF: 0.25 mm InertCap 5MS/NP (GL Science) was used as the GC column. The inlet temperature was 230 C, and column flow rate was 1.12 mL/ min. High purity helium (Air Liquide Kogyo Gas Ltd., Tokyo, Japan) was used as the carrier gas. Column temperature was maintained at 80 C for 2 min and then increased to 320 C at a rate of 15 C/min and held for 6 min. Transfer line and source temperature were 250 C and 200 C, respectively. Electron ionization was performed at 70 V. Ten scans per second were recorded over a mass range of 85e500 m/z. The GC/MS analysis data were exported in net CDF format, and peak detection and alignment were performed using MetAlign Ver. 041012 (20). AIoutput ver. 1.29 was used for peak annotation (21). Automatically annotated peaks were manually confirmed by Automated Mass Spectral Deconvolution and Identification System. LC-MS/MS analysis The lyophilized sample in 30 mL of ultrapure water (Thermo Fisher Scientific K. K., Waltham, MA, USA) was measured by ion pair LC-MS (LCMS8030 Plus: Shimadzu). L-column 2ODS (150 mm  particle size 3 mm, Chemicals Evaluation and Research Institute, Tokyo, Japan) was used as LC column. The column flow rate was 0.2 mL/min. Mobile phase A was 10 mM tributylamine with 15 mM acetate in ultrapure water, while mobile phase B was methanol. The percentage of mobile phase B was held at 0% for 1 min and raised at rates of 10%/min until 15% and held for 3 min, 7%/min until 50%, 5%/min until 100% and held for 1 min and decreased to 0% in 0.5 min. Injection volume was 1 mL, probe position was þ1.5 nm, the temperature of de-solvent line was maintained at 250 C, drying gas flow rate was 15 L/min, the temperature of heat block was maintained at 400 C and nebulizer gas flow rate was 2 L/min. Metabolites were ionized by negative ion mode. The data file from LC-MS/MS was converted to abf file via an Abf file converter (Reifycs Inc., Tokyo, Japan) and analyzed by multi reaction monitoring based probabilistic system for widely targeted metabolomics (MRMPROBS) (22). Data processing Principle component analysis was performed using the statistical analysis tool used in a previous study (23). Relative intensity of each metabolite was normalized by the internal standard included in data from GCeMS and LC-MS/MS. Auto scaling was used as the scaling method, and 1/4 root transformation was performed. Flux ratio analysis Cells at the late log phase (S. cerevisiae: OD600 ¼ 3.0 and C. utilis: OD600 ¼ 8.0) were rapidly transferred into a mixed cellulose ester membrane filter with a 47-mm diameter and 0.45-mm pore size (Toyo Roshi) by vacuum filtration and washed with distilled water. To measure alanine in the cell hydrolysate, 2.0 mg of cells were treated with 2.0 mL of 6 N HCl and hydrolyzed at 105 C for 18 h. Following filtration, 500 mL of hydrolysate was added to 10 mL of internal standard (600 mM cycloleucine) and dried via evaporation. The dry residue was dissolved in 50 mL of acetonitrile and 50 mL of N-(tertbutyldimethylsilyl)-N-methyl-trifluoroacetamide containing 1% tertbutyldimethylchlorosilane (Sigma) and incubated at 95 C for 1 h. After cooling, the supernatant was injected for GCeMS analysis. The mass isotopic distribution of ion clusters at mass to charge (m/z) ratios of [M-85], derived from alanine, was measured via GCMS QP2010 Ultra. A 30 m  0.25 mm ID: 0.25 mm DB-5MStDG (Agilent Technologies, Inc., Santa Clara, CA, USA) was used as the GC column. The column flow rate was 1.0 mL/min. High purity helium was used as the carrier gas. Injection volume was 1 mL and split ratio was 1:10. Column temperature was maintained at 142 C for 4 min and then increased to 155 C at a rate of 1 C/min, and 300 C at a rate of 10 C/min. Next, the flux ratio between glycolysis and the pentose phosphate pathway (PPP) was estimated from isotopic distribution of alanine following a previously described method (24). Labeled and non-labeled alanine from D-glucose-1-13C occur in 1:1 via glycolysis. 5/3 mol of non-labeled alanine occurred per 1 mol of D-glucose1-13C via PPP since labeled carbon was emitted as CO2. The flux ratio between glycolysis and PPP is represented by Eqs. 4 and 5. A: B ¼ X þ 5/3Y: X

(4)

1¼XþY

(5)

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

VOL. xxx, xxxx

METABOLIC COMPARISON OF CRABTREE EFFECT

A and B represent the normalized mass isotopic distribution of fragments M0 and M1 of alanine [M-85]. X and Y represent the flux ratio in glycolysis and PPP. The flux of glycolysis or PPP (mmol/gDCW/h) was calculated by following formula: specific glucose consumption rate  the flux ratio of glycolysis or PPP. Calculation of specific NADH production rate To calculate the specific NADH production rate from the specific carbon rate estimated via culture results, the following hypotheses were used: all carbon from glucose uptake is converted to cell synthesis, excreted metabolites, and CO2; molecular weight of biomass per carbon is regarded as 26.4 (25); and CO2 is emitted when glucose passes through PPP, TCA cycle, and the ethanol synthetic pathway. CO2 emitted through the ethanol synthesis pathway was calculated using the specific ethanol production rate. CO2 emitted through PPP was calculated via flux ratio analysis. Residual CO2 is emitted during the TCA cycle. 5 mol of GAP was converted from 3 mol of G6P via PPP. Next, NADH was synthesized by glycolysis and TCA cycle. 1 mol of NADH was synthesized when 1/2 mol of glucose or 3/4 mol of CO2 were consumed during glycolysis or the TCA cycle, respectively. On the other hand, 1 mol of NADH was oxidized when 1 mol of ethanol or glycerol was synthesized.

RESULTS Physiology and growth We performed aerobic culture in S. cerevisiae and C. utilis to confirm whether these yeasts exhibit the Crabtree effect (Fig. S1). Moreover, Cultivation time from 0 to 3 h was defined as lag phase. After that, depending on glucose concentration, growth phase was determined (early log phase: >9.5 g/L, middle log phase: 7.0e9.5 g/L, late log phase: 1.0e7.0 g/ L, stationary phase: <1.0 g/L). When specific rates were calculated from culture results, these rates at only middle and late log phase were constant. As expected, S. cerevisiae inhibited oxygen consumption and produced ethanol, while C. utilis did not produce any ethanol. Therefore, we confirmed that S. cerevisiae was a Crabtree positive yeast while C. utilis was a Crabtree negative yeast. The specific growth rate of S. cerevisiae (0.400  0.014 h1) was lower than that of C. utilis (0.592  0.014 h1), while the specific glucose consumption rate of S. cerevisiae (14.6  0.5 mmol/gDCW/h) was higher than that of C. utilis (6.27  0.19 mmol/gDCW/h). This indicated that the Crabtree effect does not induce cell formation but promotes excretion of ethanol from glucose as the carbon source. Consequently, cell yield of S. cerevisiae (0.183  0.002 gDCW/gglucose) was lower than that of C. utilis (0.574  0.006 gDCW/gglucose). In addition, S. cerevisiae excreted glycerol (0.88  0.05 mmol/gDCW/h), a by-product of ethanol production, as well as acetate (0.08  0.01 mmol/gDCW/h) and citrate (0.016  0.000 mmol/gDCW/h). On the other hand, C. utilis accumulated a large amount of citrate (0.034  0.004 mmol/ gDCW/h) and succinate (0.056  0.004 mmol/gDCW/h) compared to S. cerevisiae. This result suggested that C. utilis efficiently converted absorbed glucose to metabolites during the TCA cycle via glycolysis since these metabolites were synthesized during the TCA cycle. Intracellular metabolome data of S. cerevisiae and C. utilis LC-MS/MS and GCeMS were performed to measure metabolites extracted from cells at lag phase, early log phase, middle log phase, late log phase and stationary phase (Table S1). As a result, 63 and 21 metabolites were identified by LC-MS/MS and GC/MS, respectively, of which 22 common metabolites were identified by LC-MS/MS and GCeMS (Tables S2 and S3). Mainly, metabolites such as sugars, sugar phosphates, amino acids, organic acids, and cofactors were identified. Principle component analysis of metabolome data separated S. cerevisiae from C. utilis along PC1 (Fig. 1A). In culture results, these yeasts represented different specific glucose consumption rate and specific ethanol production rate (Table 1). Previous report showed that these differences depended on the Crabtree effect (26). Therefore, the separation of PC1 was caused by the Crabtree effect and metabolites which contributed to this separation were influenced

3

by the Crabtree effect. From loading data of PC1, 1,3-BPG, 3-PG/ 2 PG, PEP, 6PGA, Ru5P, citrate, isocitrate, 2-OG, fumarate, succinate, malate, ATP, AMP, and adenine may represent clues which may lead to an understanding of the different characteristics of these yeasts (Fig. 1B). In central metabolite map, these were metabolites included in glycolysis, PPP and TCA cycle as well as the number of nucleotides (Fig. 2). First, S. cerevisiae showed lower amounts of 1,3-BPG, 3PG/2 PG and PEP not only at later glycolysis, but also of 6PGA and Ru5P/Xu5P at the entrance of PPP, compared with C. utilis. Secondly, in regard to the TCA cycle, S. cerevisiae showed larger amounts of citrate and isocitrate, and smaller amount of succinate, fumarate, and malate, compared with C. utilis. Finally, the amounts of AMP and adenine were higher, but the amount of ATP was lower in S. cerevisiae, compared with C. utilis. Measurement of the branching ratio of G6P Most NADPH in cells is synthesized through PPP (26). Therefore, measuring the branching ratio between glycolysis and PPP may enhance the understanding of various characteristics of these yeasts. In this study, the branching ratio of G6P was investigated by measuring the ratio between non-labeled and C1-labeled alanine in a carbon-labeled experiment (Table 2). Results indicated that S. cerevisiae (13.5  2.0%) converted glucose uptake to PPP at a lower frequency compared with C. utilis (48.8  0.9%). In addition, PPP flux in S. cerevisiae (1.97  0.06 mmol/gDCW/h) was also lower, compared with that of C. utilis (3.06  0.09 mmol/ gDCW/h). Redox balance S. cerevisiae produced a large amount of ethanol during aerobic culture (Fig. S1). Since alcohol dehydrogenase (27) converts NADH and acetaldehyde to NADþ and ethanol, S. cerevisiae reoxidizes a large amount of NADH through ethanol production. Therefore, it is important to compare the specific rate of NADH at central metabolism in order to understand the Crabtree effect. In this study, the specific NADH production rate was calculated using the specific glucose consumption rate, PPP flux and the specific production rate of excreted metabolites (Fig. 3). Since growth phases when all specific rates were constant were only middle and late log phase (Table 1), we calculated the specific NADH production rate at late log phase. S. cerevisiae (45.0  4.7 mmol/gDCW/h) showed a higher specific NADH production rate during glycolysis and the TCA cycle, compared with C. utilis (27.1  1.2 mmol/gDCW/h). The above difference depends on the high specific glucose consumption rate. On the other hand, oxidation of NADH due to production of ethanol (18.3  0.2 mmol/gDCW/h) and glycerol (0.88  0.04 mmol/gDCW/h) occurred only in S. cerevisiae due to the Crabtree effect. Consequently, there were no substantial differences between the apparent specific NADH production rates of S. cerevisiae (25.8  4.9 mmol/gDCW/h) and C. utilis (27.1  1.2 mmol/gDCW/h). DISCUSSION Yeasts convert NADþ to NADH through glycolysis in cytosol as well as TCA cycle in mitochondria. Cytosolic NADH is reoxidized by NDE1/2 (28) and glycerol-3-phosphate shuttle (29) in respiratory chain. However, S. cerevisiae represented higher specific glucose consumption rate than C. utilis. Therefore NADþ/NADH was disrupted due to excess NADH synthesis in S. cerevisiae. To avoid NADH accumulation, glyceraldehyde 3-phosphate dehydrogenase, which converts GAP to 1,3-BPG (30), was inhibited and ethanol production was promoted by alcohol dehydrogenase which utilizes NADH as substrate. This indicated that S. cerevisiae showed a lower amount of 1,3-BPG, 3PG/2 PG and PEP than C. utilis (Fig. 2). In addition, yeasts precisely regulated the amount of NADþ and NADH as well as

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

4

IMURA ET AL.

J. BIOSCI. BIOENG.,

FIG. 1. Results of principle component analysis of metabolome data on S. cerevisiae and C. utilis. (A) Score plot (contribution rate: PC1 ¼ 31.2%, PC2 ¼ 24.8%). (B) Loading data related to PC1. (A) Circle indicates S. cerevisiae and square indicates C. utilis. The time course result is indicated by gradient from lag phase (white) to stationary phase (black).

the apparent specific NADH production rate since there were no substantial differences in these components between S. cerevisiae and C. utilis. Therefore, this result also suggested S. cerevisiae reoxidized excess NADH in cytosol by producing ethanol and glycerol. On the other hand, NADH synthesized through TCA cycle is reoxidized by mitochondrial alcohol dehydrogenase in respiratory chain (31). However, previous report proposed that mitochondrial NADH reoxidization in S. cerevisiae was not enough since it reoxidized excess NADH by converting to ethanol in fully aerobic cultivation (32). Therefore, excess NADH in mitochondria inhibited isocitrate dehydrogenase, oxoglutarate dehydrogenase and malate dehydrogenase which converted to NADH. Consequently, the amounts of citrate and isocitrate were higher, while the amounts of succinate, fumarate, and malate were lower in S. cerevisiae, compared with those in C. utilis. At the same time, S. cerevisiae exhibited low amounts of ATP compared to that of C. utilis. This

result was due to lowering of the capacity to synthesize ATP since S. cerevisiae does not utilize oxygen as the terminal electron acceptor, causing oxidative phosphorylation to be ineffective. Low oxidative phosphorylation in S. cerevisiae, inhibits the electron transfer system, thereby lowering the activity of complex II (succinate dehydrogenase); (33). Hence, the amounts of succinate, fumarate, and malate in S. cerevisiae were low compared to those of C. utilis, since the activity of isocitrate dehydrogenase, oxoglutarate dehydrogenase, and succinate dehydrogenase was low. In addition, there were no significant differences in specific NADH production rate through TCA cycle even though specific glucose consumption rate of S. cerevisiae was higher than that of C. utilis. Previous report also suggested that NADH was oxidized in S. cerevisiae through reductive branch of TCA cycle (34). These results indicated that S. cerevisiae maintained the balance between oxidative and reductive branch of TCA cycle to avoid excess NADH.

TABLE 1. Growth characteristics of S. cerevisiae and C. utilis at middle and late log phase in aerobic culture. Specific growth rate (h1) S. cerevisiae C. utilis

0.400  0.014 0.592  0.014

Specific production rate (mmol/gDCW/h) Ethanol

Acetate

Glycerol

Citrate

Succinate

18.3  0.2 n. d.

0.08  0.01 n. d.

0.88  0.05 n. d.

0.016  0.000 0.034  0.004

n. d. 0.056  0.004

Specific glucose consumption rate (mmol/gDCW/h)

Yield (gDCW/g-glucose)

14.6  0.5 6.27  0.19

0.183  0.002 0.574  0.006

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

VOL. xxx, xxxx

METABOLIC COMPARISON OF CRABTREE EFFECT

5

FIG. 2. Metabolite map depicting the lag phase, early log phase, middle log phase, late log phase and stationary phase, from left to right. Closed bars indicate S. cerevisiae and open bars indicate C. utilis. The vertical axis indicates relative intensity. All data are expressed as the mean of triplicate experiments.

Next, in regard to PPP, the amount of 6PGA and Ru5P/Xu5P were lower in S. cerevisiae than that in C. utilis. Flux ratio analysis revealed that PPP flux in S. cerevisiae was lower than that of C. utilis (Table 2). Therefore, it is assumed that S. cerevisiae showed lower specific NADPH and R5P production rates synthesized through PPP. Specific R5P production rate reflects specific growth rate since R5P is the precursor of nucleotides. Consequently, the specific growth rate of S. cerevisiae was low compared to that of C. utilis. On the

TABLE 2. Branching ratio of G6P between glycolysis and pentose phosphate pathway at and late log phase. Ratio (%)

S. cerevisiae C. utilis

Flux (mmol/gDCW/h)

Glycolysis

PPP

Glycolysis

PPP

86.5  2.0 51.1  0.9

13.5  2.0 48.8  0.9

12.6  0.4 3.21  0.10

1.97  0.06 3.06  0.09

other hand, PPP flux of S. cerevisiae was low while the amount of NADPH of S. cerevisiae was high compared to C. utilis (Fig. 2). S. cerevisiae synthesizes NADPH via other enzymes, such as aldehyde dehydrogenase which converts acetaldehyde to acetate (35). This indicated that S. cerevisiae may compensate for the lack of NADPH by using aldehyde dehydrogenase since it accumulated acetate in the medium. In addition, NADPH was utilized for cell formation (36). The amount of NADPH necessary for cell formation in C. utilis was more than S. cerevisiae due to high specific growth rate. Consequently, S. cerevisiae represented the high amount of NADPH despite of low PPP flux. Amino acids represented different amounts in each growth phase of both yeasts (Fig. 2). Therefore, they are presumed to depend on their growth phases. Since the amount of most of the amino acids did not decrease according to growth, these seem to be synthesized universally in all growth phases. On the other hand, the

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

6

IMURA ET AL.

J. BIOSCI. BIOENG.,

FIG. 3. Specific NADH production and consumption rates for S. cerevisiae and C. utilis at the late log phase.

amount of methionine, histidine and leucine at lag phase was higher than that at log phase and stationary phase. In this study, YPD medium was used for pre-cultivation. Consequently, methionine, histidine and leucine were derived from pre-culture medium which were consumed at lag phase. This indicated that batch cultures at lag phase were influenced by pre-cultures. In S. cerevisiae, NADH synthesis through catabolism especially in glycolysis greatly exceeded NADH consumption in respiratory chain. This caused NADþ/NADH imbalance and therefore it inhibited enzymes which synthesis NADH in glycolysis and TCA cycle and produced ethanol. Consequently, despite of advantage process for cell formation such as high capacity to absorb glucose, S. cerevisiae converted much of glucose to ethanol, showing low cell yield. However, this metabolism to synthesis excess NADH is attractive for production of lactate (37) and 2,3-butanediol (38,39), since NADH is necessary to produce these metabolites. Therefore, these metabolites are potential for commercial manufacturing. On the other hand, C. utilis, a Crabtree negative yeast, showed a low specific glucose consumption rate. However, C. utilis did not produce by-products such as ethanol which consume excess NADH, and metabolome analysis revealed that the amount of metabolites related to NADH production was high in C. utilis since it maintained a proper NADþ/NADH ratio compared to S. cerevisiae. Therefore, C. utilis shows potential for achieving high cell yields, due to its ability to efficiently convert absorbed glucose into biomass. High cell yield is one of the most important factors in manufacturing process. Recently, Crabtree negative yeasts superior in cell yield were the research object in bioprocess (40,41). Therefore, these yeasts might be alternative attractive hosts for S. cerevisiae. Supplementary data to this article can be found online at https://doi.org/10.1016/j.jbiosc.2019.07.007. ACKNOWLEDGMENTS This work partly financially supported by Mitsubishi Corporation Life Sciences Limited. Makoto Imura and Ryo Iwakiri are employees of Mitsubishi Corporation Life Sciences Limited. These do not alter the author’s adherence to the journal policy on sharing experimental data. The study represents a portion of the dissertation submitted by Makoto Imura to Osaka University in partial fulfillment of the requirement for his doctoral degree. References 1. Ibsen, K. H.: The Crabtree effect: a review, Cancer Res., 21, 829e841 (1961). 2. Diaz-Ruiz, R., Rigoulet, M., and Devin, A.: The Warburg and Crabtree effects: on the origin of cancer cell energy metabolism and of yeast glucose repression, Biochim. Biophys. Acta, 1807, 568e576 (2011).

3. Flikweert, M. T., van der Zanden, L., Janssen, W. M., Steensma, H. Y., van Dijken, J. P., and Pronk, J. T.: Pyruvate decarboxylase: an indispensable enzyme for growth of Saccharomyces cerevisiae on glucose, Yeast, 12, 247e257 (1996). 4. Theron, C. W., Berrios, J., Delvigne, F., and Fickers, P.: Integrating metabolic modeling and population heterogeneity analysis into optimizing recombinant protein production by Komagataella (Pichia) pastoris, Appl. Microbiol. Biotechnol., 102, 69e80 (2017). 5. Fonseca, G. G., Gombert, A. K., Heinzle, E., and Wittmann, C.: Physiology of the yeast Kluyveromyces marxianus during batch and chemostat cultures with glucose as the sole carbon source, FEMS Yeast Res., 7, 422e435 (2007). 6. Kavs cek, M., Bhutada, G., Madl, T., and Natter, K.: Optimization of lipid production with a genome-scale model of Yarrowia lipolytica, BMC Syst. Biol., 9, 72 (2015). 7. Boze, H., Moulin, G., and Galzy, P.: Production of food and fodder yeasts, Crit. Rev. Biotechnol., 12, 65e86 (1992). 8. Tamakawa, H., Ikushima, S., and Yoshida, S.: Efficient production of L-lactic acid from xylose by a recombinant Candida utilis strain, J. Biosci. Bioeng., 113, 73e75 (2012). 9. Ikushima, S., Fujii, T., Kobayashi, O., Yoshida, S., and Yoshida, A.: Genetic engineering of Candida utilis yeast for efficient production of L-lactic acid, Biosci. Biotechnol. Biochem., 73, 1818e1824 (2009). 10. Tamakawa, H., Mita, T., Yokoyama, A., Ikushima, S., and Yoshida, S.: Metabolic engineering of Candida utilis for isopropanol production, Appl. Microbiol. Biotechnol., 97, 6231e6239 (2013). 11. Iwakiri, R., Eguchi, S., Noda, Y., Adachi, H., and Yoda, K.: Isolation and structural analysis of efficient autonomously replicating sequences (ARSs) of the yeast Candida utilis, Yeast, 22, 1049e1060 (2005). 12. Iwakiri, R., Noda, Y., Adachi, H., and Yoda, K.: Isolation of the YAP1 homologue of Candida utilis and its use as an efficient selection marker, Yeast, 22, 1079e1087 (2005). 13. Iwakiri, R., Noda, Y., Adachi, H., and Yoda, K.: Isolation and characterization of promoters suitable for a multidrug-resistant marker CuYAP1 in the yeast Candida utilis, Yeast, 23, 23e34 (2006). 14. Putri, S. P., Nakayama, Y., Matsuda, F., Uchikata, T., Kobayashi, S., Matsubara, A., and Fukusaki, E.: Current metabolomics: practical applications, J. Biosci. Bioeng., 115, 579e589 (2013). 15. Teoh, S. T., Putri, S., Mukai, Y., Bamba, T., and Fukusaki, E.: A metabolomicsbased strategy for identification of gene targets for phenotype improvement and its application to 1-butanol tolerance in Saccharomyces cerevisiae, Biotechnol. Biofuels, 8, 144 (2015). 16. Ohta, E., Nakayama, Y., Mukai, Y., Bamba, T., and Fukusaki, E.: Metabolomic approach for improving ethanol stress tolerance in Saccharomyces cerevisiae, J. Biosci. Bioeng., 121, 399e405 (2015). 17. Nusantara, Putra, F. J., Putri, S. P., and Fukusaki, E.: Metabolomics-based profiling of three terminal alkene-producing Jeotgalicoccus spp. during different growth phase, J. Biosci. Bioeng., 127, 52e58 (2019). 18. Hashim, Z., Teoh, S. T., Bamba, T., and Fukusaki, E.: Construction of a metabolome library for transcription factor-related single gene mutants of Saccharomyces cerevisiae, J. Chromatogr. B. Anal. Technol. Biomed. Life. Sci., 966, 83e92 (2014). 19. Tsugawa, H., Bamba, T., Shinohara, M., Nishiumi, S., Yoshida, M., and Fukusaki, E.: Practical non-targeted gas chromatography/mass spectrometrybased metabolomics platform for metabolic phenotype analysis, J. Biosci. Bioeng., 112, 292e298 (2011). 20. Lommen, A.: Metalign: interface-driven, versatile metabolomics tool for hyphenated full-scan mass spectrometry data preprocessing, Anal. Chem., 81, 3079e3086 (2009). 21. Tsugawa, H., Tsujimoto, Y., Arita, M., Bamba, T., and Fukusaki, E.: GC/MS based metabolomics: development of a data mining system for metabolite identification by using soft independent modeling of class analogy (SIMCA), BMC Bioinformatics, 12, 131 (2011). 22. Tsugawa, H., Arita, M., Kanazawa, M., Ogiwara, A., Bamba, T., and Fukusaki, E.: MRMPROBS: a data assessment and metabolite identification tool for large-scale multiple reaction monitoring based widely targeted metabolomics, Anal. Chem., 85, 5191e5199 (2013). 23. Tsugawa, H., Cajka, T., Kind, T., Ma, Y., Higgins, B., Ikeda, K., Kanazawa, M., VanderGheynst, J., Fiehn, O., and Arita, M.: MS-DIAL: data-independent MS/ MS deconvolution for comprehensive metabolome analysis, Nat. Methods, 12, 523e526 (2015). 24. Morita, K., Nomura, Y., Ishii, J., Matsuda, F., Kondo, A., and Shimizu, H.: Heterologous expression of bacterial phosphoenol pyruvate carboxylase and Entner-Doudoroff pathway in Saccharomyces cerevisiae for improvement of isobutanol production, J. Biosci. Bioeng., 124, 263e270 (2017). 25. Lange, H. C. and Heijnen, J. J.: Statistical reconciliation of the elemental and molecular biomass composition of Saccharomyces cerevisiae, Biotechnol. Bioeng., 75, 334e344 (2001). 26. Christen, S. and Sauer, U.: Intracellular characterization of aerobic glucose metabolism in seven yeast species by 13C flux analysis and metabolomics, FEMS Yeast Res., 11, 263e272 (2011).

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007

VOL. xxx, xxxx 27. Leskovac, V., Trivi c, S., and Pericin, D.: The three zinc-containing alcohol dehydrogenases from baker’s yeast, Saccharomyces cerevisiae, FEMS Yeast Res., 2, 481e494 (2002). 28. Luttik, M. A., Overkamp, K. M., Kötter, P., de Vries, S., van Dijken, J. P., and Pronk, J. T.: The Saccharomyces cerevisiae NDE1 and NDE2 genes encode separate mitochondrial NADH dehydrogenases catalyzing the oxidation of cytosolic NADH, J. Biol. Chem., 273, 24529e24534 (1988). 29. Larsson, C., Påhlman, I. L., Ansell, R., Rigoulet, M., Adler, L., and Gustafsson, L.: The importance of the glycerol 3-phosphate shuttle during aerobic growth of Saccharomyces cerevisiae, Yeast, 14, 347e357 (1998). 30. Theobald, U., Mailinger, W., Baltes, M., Rizzi, M., and Reuss, M.: In vivo analysis of metabolic dynamics in Saccharomyces cerevisiae: I. Experimental observations, Biotechnol. Bioeng., 55, 305e316 (1997). 31. Bakker, B. M., Bro, C., Kötter, P., Luttik, M. A., van Dijken, J. P., and Pronk, J. T.: The mitochondrial alcohol dehydrogenase Adh3p is involved in a redox shuttle in Saccharomyces cerevisiae, J. Bacteriol., 182, 4730e4737 (2000). 32. Aceituno, F. F., Orellana, M., Torres, J., Mendoza, S., Slater, A. W., Melo, F., and Agosin, E.: Oxygen response of the wine yeast Saccharomyces cerevisiae EC1118 grown under carbon-sufficient, nitrogen-limited enological conditions, Appl. Environ. Microbiol., 78, 8340e8352 (2012). 33. Takeda, M.: Glucose-induced inactivation of mitochondrial enzymes in the yeast Saccharomyces cerevisiae, Biochem. J., 198, 281e287 (1981). 34. Nissen, T. L., Schulze, U., Nielsen, J., and Villadsen, J.: Flux distributions in anaerobic, glucose-limited continuous cultures of Saccharomyces cerevisiae, Microbiology, 143, 203e218 (1997).

METABOLIC COMPARISON OF CRABTREE EFFECT

7

35. Grabowska, D. and Chelstowska, A.: The ALD6 gene product is indispensable for providing NADPH in yeast cells lacking glucose-6-phosphate dehydrogenase activity, J. Biol. Chem., 278, 13984e13988 (2003). 36. Matsuda, F., Furusawa, C., Kondo, T., Ishii, J., Shimizu, H., and Kondo, A.: Engineering strategy of yeast metabolism for higher alcohol production, Microb. Cell Fact., 10, 70 (2011). 37. Baek, S. H., Kwon, E. Y., Bae, S. J., Cho, B. R., Kim, S. Y., and Hahn, J. S.: Improvement of d-lactic acid production in Saccharomyces cerevisiae under acidic conditions by evolutionary and rational metabolic engineering, Biotechnol. J., 12, 1700015 (2017). 38. Ishii, J., Morita, K., Ida, K., Kato, H., Kinoshita, S., Hataya, S., Shimizu, H., Kondo, A., and Matsuda, F.: A pyruvate carbon flux tugging strategy for increasing 2,3-butanediol production and reducing ethanol subgeneration in the yeast Saccharomyces cerevisiae, Biotechnol. Biofuels, 11, 180 (2018). 39. Kim, J. W., Kim, J., Seo, S. O., Kim, K. H., Jin, Y. S., and Seo, J. H.: Enhanced production of 2,3-butanediol by engineered Saccharomyces cerevisiae through fine-tuning of pyruvate decarboxylase and NADH oxidase activities, Biotechnol. Biofuels, 9, 265 (2016). 40. Qiao, K., Wasylenko, T. M., Zhou, K., Xu, P., and Stephanopoulos, G.: Lipid production in Yarrowia lipolytica is maximized by engineering cytosolic redox metabolism, Nat. Biotechnol., 35, 173e177 (2017). 41. Shang, T., Si, D., Zhang, D., Liu, X., Zhao, L., Hu, C., Fu, Y., and Zhang, R.: Enhancement of thermoalkaliphilic xylanase production by Pichia pastoris through novel fed-batch strategy in high cell-density fermentation, BMC Biotechnol., 17, 55 (2017).

Please cite this article as: Imura, M et al., Comparison of metabolic profiles of yeasts based on the difference of the Crabtree positive and negative, J. Biosci. Bioeng., https://doi.org/10.1016/j.jbiosc.2019.07.007